AWS Summit Washington, DC 2025 | Tim Tutt, Snowflake

Clips
News
More from AWS Summit Washington, DC 2025

Timothy Tutt

GM of Public Sector

Snowflake

Inside the AI shift in government: 11 perspectives on public sector ops transformation

Public sector transformation is entering a critical phase, driven by the rise of agentic AI and digital labor that doesn’t just automate, it adapts. But the biggest obstacle isn’t the technology. It’s the systems, silos and culture built for a slower, pre-AI world.TheCUBE’s John Furrier discusses public sector transformation.What’s really holding back progress in the public sector isn’t a lack of innovation — it’s the difficulty of adoption. From aging infrastructure to deeply rooted bureaucracies, systemic inertia remains the biggest obstacle. Meanwhile, AI is rapidly reshaping the role of government, raising the stakes for timely transformation, according to theCUBE’s John Furrier

play_circle_outline Integration and modernization with tooling for faster innovation cycles

play_circle_outline Maximizing Security with AI and Cloud Innovation: Enhancing Trust, Processes, and Productivity

Info
Transcript

Tim Tutt, Snowflake

Timothy Tutt

GM of Public Sector Snowflake

Tim Tutt, general manager of public sector at Snowflake Inc., joins theCUBE’s John Furrier and Scott Hebner at the AWS Summit Washington, DC event to discuss the evolving role of AI and cloud infrastructure in government. The conversation highlights the importance of public-private collaboration and how modern architectures are reshaping data strategies across federal agencies.

Tutt shares how Snowflake approaches secure data management through FedRAMP certifications and IL5 readiness, enabling sensitive workloads for agencies such as the Department of... Read more

explore Keep Exploring

What impact does the fast integration cycles and modernization in the tech circles have on the capabilities and operations of Snowflake for its customers, including the Department of Defense and national security? add

What are some ways that GenAI and other AI capabilities can help enhance a company's security posture? add

bolt Powered by CUBE AI

Tim Tutt, Snowflake

search

>> Hello, welcomes with theCUBE. We are here on the ground floor at AWS DC Summit. I'm John Furrier, host of theCUBE. Scott Hebner, industry analyst with theCUBE Research. This is where all the action's happening in cloud and AI for public sector and also a little commercial because of the accelerated public-private partnerships. And of course, the growth of AI is transforming architectures in computing and software, certainly with agents and AI is going to change the landscape at the business logic and government, that's going to be hard, but it's happening very, very fast. Tim Tutt is here, vice president of U.S. Public Sector at Snowflake. Tim, great to have you on theCUBE. Good to see you.

>> John, pleasure to see you as well, Scott. Thank you both for having me here.

Scott Hebner

>>

>> So first, congratulations on a great event for the Snowflake Summit last week in San Francisco, we had theCUBE coverage there, team coverage, a lot of great stuff. But the big takeaway was data has to be available, it's got to be interoperable, it's got to be fast, low latency because AI demands it, the agents demand it. Your history in the government public sector, you've seen the movie, you've seen change and transformation at many levels. So the first question is, where are we now on this transformation? Scope the transformational side? I mean, you've been through the cloud with CIA. What's the scope now?

>> Absolutely. Yeah. It's interesting because what we've started to shift to is instead of moving your data to where it needs to meet the capabilities, we're now moving the capabilities to your data. I think that's one of the biggest things that I'm seeing that massive shift in is, hey, everyone wants to move to the cloud. It's taken a long time for everyone to get there. Within the national security space in particular, there's always a little bit of concern about, well, hey, what's safe? What can I do? Is this going to be okay for my data? Am I going to be able to get the value out of it? But over the last, let's call it 10 years, it's been a massive transformation where now everyone's clamoring to get there and they're really just looking to figure out, hey, what can I get out of this for myself -

>> , not my budget.

>> Yes, absolutely.

>> That's your department, you swim lanes.

>> Oh, yeah.

>> Okay, so talk about just scope, also Snowflake's position in public sector certifications levels. Give us the numbers and facts.

>> Absolutely. So we've got FedRAMP high, FedRAMP moderate. We just received our IL5 certification, had IL4 for a little bit now, which is a huge momentum shift for us because now that it opens up the aperture for us to really dive into Department of Defense space, helping to kind of target the controlled unclassified data. We can start to really look at things like force readiness. We can look at, hey, how are we analyzing financial data or even partnerships other with other countries in terms of data sharing as we're starting to plan a little bit more globally, which is a very exciting move for us here at Snowflake.

>> On the DOD side, one of the things I covered during the JEDI contract drama was the project was actually really legit. The tactical edge is very data-driven and a lot of the deaths were friendly fire. So a lot of optimization with the data. Are we there yet? Where's the progress bar on this data portability, capabilities to the edge? Are we in much better place now than we were then?

>> We're still in a little bit of a slow hold there. I would say we're getting there slowly. But the tactical edge, you're talking about interoperability between the cloud and low latency environments in these areas where I don't have a direct network connection, how do I do the operation directly in my mission area? Now, the cloud helps enable that, especially when you're deployed in these regions that, hey, it's easily accessible for you. And when you can push all the processing power off to the cloud, it makes it a whole lot easier for your analysts, your operators on the ground to get that value from their data very rapidly.

Scott Hebner

>> So it sounds like there's been a lot of progress of getting the data infrastructure. And the old adage, there's no AI without an IA, information architecture.

>> That's exactly right.

Scott Hebner

>> So obviously now this rapid transformation to build AI capabilities, which eventually is going to turn into agents on top of all that data and cloud infrastructure. What are the unique characteristics that the defense organizations have to think about as compared perhaps to other parts of the government or even the private sector?

>> Yeah, there's a different level of sensitivity of the data that you're dealing with in general. I think, listen, it's all the same general challenges. It's how do we analyze this data? How do we get value from it? How do we make the right decisions as we're moving? But the sensitivity of that data has direct impacts on force protection. So as we're moving out in the field, we don't want our adversary seeing, hey, here's the movements that they're going to make, here's what they're targeting because they can also filter in misinformation. So it becomes immensely important to make sure that the data that you're munging in is accurate data. Otherwise, you're going to wind up in a space where, hey, you're acting on inaccurate information that puts our forces in a bad position overall.

Scott Hebner

>> I got two very specific questions. One is around accountability. So I think that's one of the areas where you look at a lot of the agentic solutions out there from a workflow perspective, the credentials of these agents, who are they accountable to? They got to be accountable to human beings. What can they access? How do you track what they do? Because acting like digital workers on behalf of a person. So I'm curious how in that kind of super sensitive environment, it seems to me that's a really important need. So let's do that one first, then I got to a quick follow up on that.

>> Sure. No, so it's immensely important, and I would say this, we're not there yet on the DOD national security side. There's still a lot of policy and process work that needs to happen to ensure that are secure at the end of the day before we start enabling it. One of the key principles is always going to be human in the loop with these agentic solutions as we start pushing these out. Because as you start to give these agents more credentials, you actually open up another vector of attack. Hey, now I can go and attack that agent and see if I can funnel out information. Or, is that agent even feeding me the right information, the right decisions as we're moving forward here?

Scott Hebner

>> We're moving on from this world of AI doing correlations and pattern recognition, it's all probabilistic statistics, right?

>> Absolutely.

Scott Hebner

>> With the agents, you start to incorporate reasoning and planning and decision intelligence. So the trust factor becomes critical. Not just security, but also compliance and just humans aren't going to use it unless they trust it. That must be another big challenge, particularly in the government, right? In the defense.

>> Absolutely. Look, when you're talking about operational planning, those things matter a lot. But there are other applications from a business perspective where this can be absolutely beneficial. How do I help enable our business processes from a finance, from a HR perspective, the back office type things that help enable all of these operations to even run in the first place? One of the key areas that I really start to think about where agentic solutions can become immensely beneficial is supply chain logistics. This is an area where we can start to see massive change in the relatively near term. It's going to have some impacts, but as long as you're keeping that human in the loop, you can manage and mitigate the risk.

>> Tim, I want to get on that capabilities point. You just mentioned a lot of use cases, a lot of supply chain efficiencies are all laid out there. So how do you get there? And the question I want to ask you is, the hottest conversation in the tech circles in San Francisco, New York, DC from the Alpha Geeks I call them, is tooling and cogeneration. So we've seen probably the fastest kind of tool wave because it's easier to build tools now than ever before. Which then, if you believe that to be true, which kind of is, integration is easier. So integration becomes a very key thing. So we're seeing for the first time very fast integration cycles, which also lead to migration and modernization. So they're all systemically tied together. So how does that impact some of the things that you guys are doing at Snowflake, for your customers and the Department of Defense and national security? Because is now you can start integrating in a new fabric architecture to break down those old legal architectures that were predicated by contracts or budgets. Now you're just like, whoa, I got unfettered access to horizontally scalable data or integrate into a tool.

>> Absolutely. That's one of the things that I love about the Snowflake AI Data Cloud. We are a place for that integration and what you're starting to see more out of Snowflake writ large, and also especially us in the U.S. public sector, is a more heavy focus on how do we get those integration partners to come to the data. If we can get all of the data in Snowflake, you have every capability that you need and it's really simple and easy for you to integrate and leverage LLMs, leverage the agentic solution, machine learning tooling. You can take a containerized application, bring it to Snowflake and it runs with access to your data, all the governance and security in place to help manage across the board.

>> Are people skeptical? I mean, because it sounds too good to be true.

>> I think everyone's always skeptical until they start using it. And the thing I love about Snowflake is, it's actually that simple. We can get you up and running in a matter of hours, and once you start to see that power, it changes the dynamic pretty dramatically.

Scott Hebner

>> How is this solution with Snowflake helping them balance... I think on one end of the spectrum is you got a lot of compliance, you've got more data being generated, you've got agents now doing things, so there's a lot of sensitive things that have to be managed. Other end is innovation cycles with AI are moving at warp speed. It's too easy to fall behind.

>> Absolutely.

Scott Hebner

>> And it's not like cloud and client server and mobile where you could sit on the sidelines a little bit to see what shapes out. You got to get in there and play the game, right? So how do you balance those two worlds and how is what you just outlined going to help speed things up?

>> Well, if you look at the way Snowflake is set up, we are rapidly innovating. We're adding new capabilities over and over on a weekly basis. I mean, you look at how fast we integrate some of the newer models coming to bear inside of the Snowflake environment. Now, me as a consumer, I can actually iterate faster because I have all the capabilities I need where my data already lives. I don't have to move it somewhere else. I don't have to go through a whole another security architecture. All the capabilities I need are where my data lives.

Scott Hebner

>> And now it's Impact 5 compliant, therefore there's less sensitivity and you can move faster.

>> Absolutely. That's exactly right.

Scott Hebner

>> That's basically what you've done with AWS.

Scott Hebner

>> Yep.

>> All right, the question, the next one is, okay, the plumbing's getting worked on, the culture's changing, well, changing slowly-

>> Slowly.

>> But fast for government, right? I mean, would you say fast for government?

>> I had say it's faster for government than I've ever seen before.

>> Okay. Fast.

>> Yes.

>> It's not slow, not like what I would call moving glaciers. But, let's move beyond that. As you get to the application impact with a business logic, as you mentioned, tactical edge, that's an edge piece. There's so many applications in government that have all that business domain expertise. Where promise of AI shines is in the domain expertise process, workflows the higher end of the stack. What's the application layer impact going to be from the agentic data wave coming?

>> Again, I will note, yes, we're moving faster than we've ever seen it before, but it's going to be still a little slow because core piece of is you have to have that human in the loop. That agentic wave is going to help to make sure you're able to integrate those applications, make sure your analysts can engage in a faster way. We're really talking about streamlining processes at the end of the day, how do I help get you there faster? Not how do I help replace you? Because if we start replacing the humans, it opens up a significant -

>> So you're saying productivity and efficiency, low-hanging fruit now, people are knocking those down first because the hard problem is the infra and the data, or not?

>> It's less the infra and the data. It's more the policy and the security around it. So yes, the productivity, let's enhance those. But now how do I trust? Is it one of the key things that I'm getting the right results out of these solutions at the end of the day? And the only way you're going to trust it is if you've got that human in the loop going back and forth saying, this is right, this is wrong. How do we enhance these models? And you have to make sure you have the right human in the loop because if you train it wrong, you're never going to get the right answer.

>> I love having you on theCUBE because you got the computer science, you have the history, and also you're in great company at Snowflake. We interviewed a lot of CIOs, CISOs at big banks. So big banks, like the big banks, tier one, and government have a lot of similarities. Size, security requirements and trust. Now different levels of needs. So I was talking to a CISO and they said, we have such a high security cyber resilience bar that the security bar is so high, our resilience is so high. And she said, well, we look at GenAI as just another app review. Okay, now the question to you is, okay, the resilience level to execute and security is high. How would you see GenAI threading the needle to get into production, that path?

>> So it's interesting because from a security perspective, you can actually leverage GenAI and other AI capabilities to help enhance your security posture overall. I think threading that needle is really about finding that balance between, hey, what are my true risks? What are the things that I care about? What can I allow? In security, we've always historically had this issue of, hey, say no first. But security really should be an enabler. How do I help you get there? It's like when I go and talk to my lawyers, Hey, help me get to a yes instead of a no. Especially on the government side, you see this all the time. Your job is to be an enabler. GenAI can help you be that enabler. The rest of the AI capabilities and ecosystem can help be those enablers by automating some of the processes that you're used to seeing in security over and over again. And once you start to build that trust, going to take time, again, it becomes a productivity enabler across.

>> Final question for you. Obviously you learned a lot at the CIA, actually a rebel organization considered by its standards. You guys were really first into the cloud. Okay, so now we're at a whole other era. What kinds of advice would you give other leaders in organizations that are looking to be what CIA was for the cloud? Because there's a lot of opportunities, a lot of people are moving into the cloud and beyond. With AI, there's that similar leap of faith. Will the parachute open? Will it work?

>> Look, it's all about process and innovation. I need to have an innovative mindset and I need to have the processes to make sure that I can move through those cycles quickly. We've historically seen blockers like contractual processes and things like that. Those are not technology blockers. You've got people that can actually start engaging with these capabilities in un-classed arenas and then say, hey, how do I bring them into my mission? Let's start with something simple. Let's look at commercially available information or publicly available information. What can I get value out of there? And now I start to build that trust. And now I can start to say, let's bring this in a little bit closer to my mission core focus area.

>> Scott, it's the same people, process, technology, change management.

>> Always.

>> Tim, thank you for coming on theCUBE. Really appreciate you taking the time from your busy schedule.

>> Yep. John, Scott, I really appreciate it and thank you for having me here.

>> All right, cool. This is theCUBE. We're bringing all the action here. Again, public sectors transforming commercial private partnerships coming together. And again, security, AI. At the end of the day, it's a people problem. Bringing all the action here, I'm John Furrier with Scott Hebner. Thanks for watching.